Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Subject: Re: [PATCH 1/6] LoongArch: Get frame info in unwind_start when regs
 is not supported
To:     Qing Zhang <zhangqing@loongson.cn>,
        Huacai Chen <chenhuacai@kernel.org>,
        WANG Xuerui <kernel@xen0n.name>
Cc:     loongarch@lists.linux.dev, linux-kernel@vger.kernel.org,
        Steven Rostedt <rostedt@goodmis.org>,
        Masami Hiramatsu <mhiramat@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>
References: <20221215040141.18610-1-hejinyang@loongson.cn>
 <20221215040141.18610-2-hejinyang@loongson.cn>
 <d3c5aa06-0cc3-b1a4-18b9-858e8f1216de@loongson.cn>
From:   Jinyang He <hejinyang@loongson.cn>
Message-ID: <e4d39bdb-827a-8a9c-0aeb-fcc919743826@loongson.cn>
Date:   Thu, 15 Dec 2022 13:55:40 +0800
User-Agent: Mozilla/5.0 (X11; Linux loongarch64; rv:68.0) Gecko/20100101
 Thunderbird/68.7.0
MIME-Version: 1.0
In-Reply-To: <d3c5aa06-0cc3-b1a4-18b9-858e8f1216de@loongson.cn>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-US
Precedence: bulk

On 2022-12-15 13:24, Qing Zhang wrote:

> Hi, Jinyang
>
> On 2022/12/15 下午12:01, Jinyang He wrote:
>> At unwind_start, it is better to try to get its frame info even we not
>> support regs rather than get them outside. So that we can simply use
>> unwind_{start, next_frame, done} outside.
>>
>> Signed-off-by: Jinyang He <hejinyang@loongson.cn>
>> ---
>>   arch/loongarch/kernel/process.c         | 12 +++---------
>>   arch/loongarch/kernel/unwind_guess.c    |  6 ++++++
>>   arch/loongarch/kernel/unwind_prologue.c | 16 +++++++++++++---
>>   3 files changed, 22 insertions(+), 12 deletions(-)
>>
>> diff --git a/arch/loongarch/kernel/process.c 
>> b/arch/loongarch/kernel/process.c
>> index 502b8b950057..6ef45174ad35 100644
>> --- a/arch/loongarch/kernel/process.c
>> +++ b/arch/loongarch/kernel/process.c
>> @@ -197,20 +197,14 @@ int copy_thread(struct task_struct *p, const 
>> struct kernel_clone_args *args)
>>     unsigned long __get_wchan(struct task_struct *task)
>>   {
>> -    unsigned long pc;
>> +    unsigned long pc = 0;
>>       struct unwind_state state;
>>         if (!try_get_task_stack(task))
>>           return 0;
>>   -    unwind_start(&state, task, NULL);
>> -    state.sp = thread_saved_fp(task);
>> -    get_stack_info(state.sp, state.task, &state.stack_info);
>> -    state.pc = thread_saved_ra(task);
>> -#ifdef CONFIG_UNWINDER_PROLOGUE
>> -    state.type = UNWINDER_PROLOGUE;
>> -#endif
>> -    for (; !unwind_done(&state); unwind_next_frame(&state)) {
>> +    for (unwind_start(&state, task, NULL); !unwind_done(&state);
>> +         unwind_next_frame(&state)) {
>>           pc = unwind_get_return_address(&state);
>>           if (!pc)
>>               break;
>> diff --git a/arch/loongarch/kernel/unwind_guess.c 
>> b/arch/loongarch/kernel/unwind_guess.c
>> index e2d2e4f3001f..e03864511582 100644
>> --- a/arch/loongarch/kernel/unwind_guess.c
>> +++ b/arch/loongarch/kernel/unwind_guess.c
>> @@ -26,6 +26,12 @@ void unwind_start(struct unwind_state *state, 
>> struct task_struct *task,
>>       if (regs) {
>>           state->sp = regs->regs[3];
>>           state->pc = regs->csr_era;
>> +    } else if (task == current) {
>> +        state->sp = (unsigned long)__builtin_frame_address(0);
>> +        state->pc = (unsigned long)__builtin_return_address(0);
>> +    } else {
>> +        state->sp = thread_saved_fp(task);
>> +        state->pc = thread_saved_ra(task);
> The case for tasks is already handled in stacktrace.c, which is 
> reusable code:
> void arch_stack_walk(stack_trace_consume_fn consume_entry, void *cookie,
>              struct task_struct *task, struct pt_regs *regs)
> {
> ...
>
>     if (task == current) {
>         regs->regs[3] = (unsigned long)__builtin_frame_address(0);
>         regs->csr_era = (unsigned long)__builtin_return_address(0);
>     } else {
>         regs->regs[3] = thread_saved_fp(task);
>         regs->csr_era = thread_saved_ra(task);
>     }
> ...
>     for (unwind_start(&state, task, regs);
>           !unwind_done(&state); unwind_next_frame(&state)) {
>         addr = unwind_get_return_address(&state);
> ...
> }
>
>>         state->task = task;
>> diff --git a/arch/loongarch/kernel/unwind_prologue.c 
>> b/arch/loongarch/kernel/unwind_prologue.c
>> index 0f8d1451ebb8..9d51ea37782e 100644
>> --- a/arch/loongarch/kernel/unwind_prologue.c
>> +++ b/arch/loongarch/kernel/unwind_prologue.c
>> @@ -141,12 +141,22 @@ void unwind_start(struct unwind_state *state, 
>> struct task_struct *task,
>>               struct pt_regs *regs)
>>   {
>>       memset(state, 0, sizeof(*state));
>> +    state->type = UNWINDER_PROLOGUE;
>>   -    if (regs && __kernel_text_address(regs->csr_era)) {
>> -        state->pc = regs->csr_era;
>> +    if (regs) {
>>           state->sp = regs->regs[3];
>> +        state->pc = regs->csr_era;
>>           state->ra = regs->regs[1];
>> -        state->type = UNWINDER_PROLOGUE;
>> +        if (!__kernel_text_address(state->pc))
>> +            state->type = UNWINDER_GUESS;
>> +    } else if (task == current) {
>> +        state->sp = (unsigned long)__builtin_frame_address(0);
>> +        state->pc = (unsigned long)__builtin_return_address(0);
>> +        state->ra = 0;
>> +    } else {
>> +        state->sp = thread_saved_fp(task);
>> +        state->pc = thread_saved_ra(task);
>> +        state->ra = 0;
>>       }
> also...
>>       state->task = task;
>>
That looks similar, but stacktrace actually provides the regs parameter.
Unwind_{start, next_frame, get_return_address} is API to others, such the
GPL module. Btw, if we want to drop the redundant codes in arch_stack_walk
and keep same result, use "cookie->skip += 1", I think, but beyond this 
series.


Thanks,

Jinyang